Frequent sleep medication is associated with an increased risk of dementia in White older adults


The effect of sleep medications on cognition in older adults is controversial, possibly dependent upon sleep quality, and may differ by race. 

According to a recent scientific study led by the University of California, San Francisco, sleep medications, including benzodiazepines, antidepressants, and Ambien, may increase the risk of dementia in white individuals.

. . . .

Benzodiazepines (BZDs) are commonly prescribed for anxiety and agitations, which are early symptoms of Alzheimer’s disease and related dementias (ADRD). It is unclear whether BZDs causally affect ADRD risk or are prescribed in response to early symptoms of dementia.

The study findings were published in the Journal of Alzheimer’s Disease.

The researchers have examined 3,068 community-dwelling older adults (aged 74.1±2.9 years, 41.7% Black, 51.5% female) without dementia.

Sleep medication use was recorded three times by asking “Do you take sleeping pills or other medications to help you sleep?” with the response options: “Never (0)”, “Rarely (≤1/month)”, “Sometimes (2–4/month)”, “Often (5–15/month)”, or “Almost Always (16–30/month)”.

Incident dementia was defined using hospitalization records, dementia medication prescription or clinically significant decline in global cognition. Results:138 (7.71%) of Whites and 34 (2.66%) of Blacks reported taking sleep medications “often or almost always”.

Whites were almost twice as likely to take all prescription hypnotics. 617 participants developed dementia over the follow-up. After adjustment for all covariates, participants who reported taking sleep medications ≥ 5/month versus ≤1/month were significantly more likely to develop dementia, and the association was only observed among Whites (HR = 1.79,1.21–2.66) but not Blacks (HR = 0.84,0.38–1.83); p for interaction = 0.048.

Further adjustment for nighttime sleep did not appreciably alter the results. The association was similar for the cumulative frequency of sleep medication use and remained after introducing a time lag of 3 years.


Frequent sleep medication use was associated with an increased risk of dementia in White older adults. Further research is needed to determine underlying mechanisms.

Pharmacology of Benzodiazepines
Benzodiazepines, which contain one or more 6-carbon benzene rings, a 7-carbon diazepine ring, and various substituents, bind to benzodiazepine receptors on the gamma aminobutyric acid (GABA) receptor complex. Understanding the nature of benzodiazepine receptors and their effect on GABAergic transmission will simplify an appreciation of the diverse actions of benzodiazepines and other medications that act on benzodiazepine receptors.

Benzodiazepine Receptors
GABA is the major inhibitory neurotransmitter in the central nervous system (CNS). The balance between inhibitory GABAergic, inhibitory glycinergic, and excitatory glutaminergic systems is fine-tuned to set the overall state of arousal of the CNS. GABA interacts with the GABA heteroreceptor, a member of the nicotinoid superfamily that is allosterically linked to membrane receptors for benzodiazepines and several other agonists to form a pentameric, GABA gated, chloride ion channel [19, 21]. Interaction of the postsynaptic GABAA receptor with GABA alters the conformation of the GABA receptor, opening the chloride channel and allowing more negative charges into the cytosol, which hyperpolarizes (and reduces excitability of) neurons that regulate vigilance, emotions, cognition, and muscle tension [22].

The benzodiazepine receptor consists of two α, two β, and one γ subunit allosterically linked to the GABAA receptor. Benzodiazepines bind to an extracellular site at the α/γ interface; there is a low-affinity benzodiazepine-binding site in the transmembrane region of the β/α interface, which is also a binding site for the anesthetics propofol and etomidate [21]. Occupancy of the benzodiazepine receptor by a benzodiazepine does not affect the chloride ion channel directly but changes the conformation of the GABA receptor to increase its affinity for GABA, resulting in more chloride ion influx and more neuronal hyperpolarization per GABA molecule. A number of other medications and substances act directly or indirectly in various ways to alter activity of the GABA-benzodiazepine receptor complex. For example, picrotoxin, a convulsant, binds directly to the chloride channel and reduces chloride channel openings to induce seizures. Alcohol and barbiturates have direct effects on GABA receptors and chloride channels (possibly at the picrotoxin binding site), reducing cortical and limbic neuronal activity. Substances that are cross-reactive with benzodiazepines, such as anesthetics, barbiturates, neurosteroids, and alcohol, bind to different sites on the GABAA-benzodiazepine receptor complex.

Three major benzodiazepine receptor subtypes have been described [23]. Benzodiazepine-1 receptors are located throughout the CNS, particularly the reticular activating system, cerebellum, and cerebral cortex, mediating sleep and anti-anxiety effects. Benzodiazepine-2 receptors, which are located in the cortex, hippocampus, striatum, and spinal cord and on pyramidal neurons, mediate anxiolysis, muscle relaxation, CNS depression, sedation, and psychomotor impairment and contribute to the anticonvulsant effect. The benzodiazepine-3 receptor, also referred to as the translocator protein (TSPO), has a structure and function distinct from that of central (i.e., benzodiazepine-1 and benzodiazepine-2) receptors and, therefore, is sometimes called the peripheral benzodiazepine receptor or PBR. PBRs are found on glial and other brain cells as well as throughout the body and may mediate cognitive side effects, physical dependence, tolerance, and withdrawal [23].

Benzodiazepine Receptor Ligands
Compounds that bind to the benzodiazepine receptor can have five different actions. Nonselective agonists (e.g., benzodiazepines) act on all benzodiazepine receptors to increase GABA receptor affinity, increasing the frequency of chloride channel openings and hyperpolarizing all neurons with benzodiazepine receptors. Selective agonists (e.g., zolpidem), which generally do not have a benzodiazepine structure, exert the same effect, but primarily at the benzodiazepine-1 receptor. Avoiding actions on benzodiazepine-2 and -3 receptors reduces the risk of adverse effects, such as psychomotor impairment, tolerance, and syndromes associated with drug discontinuation, although these still occur. Nonselective partial agonists (e.g., zopiclone) bind to all benzodiazepine receptors, but they are less effective than full agonists, resulting in a less pronounced therapeutic effect than full agonists, but with less intense adverse effects. Inverse agonists (e.g., beta carbolines) occupy the benzodiazepine receptor and alter its conformation in a manner that reduces GABA receptor affinity, decreasing the frequency of chloride channel openings, and depolarizing neurons. Inverse agonists produce arousal and anxiety and sometimes cause seizures and psychosis. Antagonists (e.g., flumazenil) occupy a receptor without affecting its activity, thereby blocking the action of agonists, partial agonists, and inverse agonists to reverse their effects.

Endogenous and exogenous benzodiazepine receptor ligands called endozepines act on benzodiazepine receptors to participate in the regulation of arousal, neurosteroid biosynthesis, neuropeptide expression, neurogenesis, and pro-inflammatory cytokine secretion [24]. An 86-amino acid endozepine called diazepam binding inhibitor (DBI) because it displaces labeled diazepam from brain binding sites, is located in astroglial cells and peripheral organs [24]. DBI, also called acyl-CoA-binding protein (ACBP), acts as a benzodiazepine receptor inverse agonist, promoting arousal. The function of DBI/ACBP appears to be to fine tune the set point of CNS arousal in conjunction with other inhibitory and excitatory influences. The clinical importance of this phenomenon is that the chronic presence of benzodiazepines downregulates benzodiazepine receptors and upregulates DBI to maintain the baseline level of arousal, the latter effect being overridden by the large number of benzodiazepine molecules compared to DBI molecules. Rapid reduction of the number of benzodiazepine molecules with discontinuation exposes the decreased number of benzodiazepine receptors to a prominent inverse agonist effect, resulting in excessive arousal that is manifested as withdrawal as discussed below.

Pharmacokinetics of Benzodiazepines
All benzodiazepines except clorazepate are completely absorbed after oral administration and reach peak serum levels within 30 min to 2 h. Metabolism of clorazepate in the stomach converts it to desmethyldiazepam (nordiazepam), which is then completely absorbed. Intramuscular (IM) absorption of benzodiazepines other than lorazepam and midazolam is slower than oral absorption. The onset of action is nearly immediate with intravenous administration of high-potency benzodiazepines such as midazolam and diazepam.

Benzodiazepine metabolism varies [20, 25, 26]. Oxazepam and lorazepam are conjugated directly by glucuronidation to a water-soluble form and are excreted without any intermediate metabolites. Most other benzodiazepines are oxidized by cytochrome P450 (CYP) 3A4 and 2C19, frequently to active metabolites. These metabolites may then be hydroxylated to another active metabolite. For example, diazepam, chlordiazepoxide, and clorazepate are oxidized to desmethyldiazepam, which has an elimination half-life of 120 h and which, in turn, is hydroxylated to oxazepam. These products then undergo glucuronidation to water soluble inactive metabolites. Flurazepam, a benzodiazepine used as a hypnotic that has a short elimination half-life, has an active metabolite (desalkylflurazepam) with a half-life greater than 100 h. The duration of action of a benzodiazepine, therefore, may not correspond to the half-life of the parent drug.

Concentrations of benzodiazepines in the CSF are equal to plasma concentrations of the free drug. Correlations have not been demonstrated between serum levels of benzodiazepines and clinical response, with the possible exception of a correlation between levels of alprazolam and both adverse effects and efficacy for panic attacks, with response more likely at plasma concentrations around 20–40 ng/mL, corresponding to doses of 1.5–6.0 mg per day. However, such levels are rarely measured in clinical practice, and levels would be expected to fluctuate substantially between doses, making the optimal time to measure the trough level difficult to identify. Whereas serum levels are not likely to predict response to any benzodiazepine, they identify benzodiazepines in toxicology screening, and undetectable levels in an agitated patient who has been taking benzodiazepines raise the suspicion of a discontinuation syndrome.

All benzodiazepines are lipid-soluble, but the degree of lipid solubility varies considerably between drugs. Benzodiazepines and their active metabolites are 70–99% protein bound, the extent of binding being proportional to their lipid solubility. Benzodiazepines follow a 2-compartment distribution, which involves a rapid central compartment phase followed by a redistribution phase to adipose tissue that ultimately determines duration of action [25]. As a result, a component of the long half-life with chronic treatment of some highly lipophilic benzodiazepines, such as diazepam, is attributable to extensive storage in adipose tissue [25]. In contrast, a single dose of a very lipophilic benzodiazepine, including diazepam, redistributes rapidly to adipose tissue from CNS receptors, leading to a short duration of action despite the long half-life, while less lipophilic benzodiazepines like lorazepam remain at therapeutic concentrations at GABAA receptors longer after a single dose because they do not redistribute as quickly [25]. The rapid onset of action of benzodiazepines that are high in lipid solubility and potency can produce quick relief of acute anxiety, but it can be disturbing to hypervigilant patients, and with chronic treatment with a short half-life agent like alprazolam, it can lead to alternations of rapid onset both of reward and distress, promoting increasing the dose, which only intensifies symptoms when brain levels of the next dose decrease faster in response to the higher CNS-blood concentration gradient [27]. Repeated dosage increases to chase interdose withdrawal and rebound can lead to excessive use compared with less lipid-soluble, less potent medications, such as chlordiazepoxide [28].

Nonbenzodiazepine Agonists of the Benzodiazepine Receptor
A group of medications called z-drugs do not have a benzodiazepine structure in that they lack a diazepine ring. Although it is technically accurate to describe these medications as nonbenzodiazepines, they are benzodiazepine receptor agonists, and their clinical effect can be blocked by the benzodiazepine receptor antagonist flumazenil. Zolpidem, an imidazopyridine, and zaleplon, a pyrazolopyrimidine, are selective for the benzodiazepine-1 receptor and have no significant effect on other benzodiazepine receptor subtypes at usual doses; at higher doses, they are less selective. Selectivity for the benzodiazepine-1 receptor results in hypnotic and some anxiolytic action without muscle relaxant or anticonvulsant effects. Lack of effects on benzodiazepine-3 receptors appears to result in a lower incidence of withdrawal and rebound symptoms, although these may occur, especially at higher doses and with more chronic treatment. Zopiclone is a nonselective benzodiazepine receptor partial agonist that is in widespread use in Europe and other countries but is not available in the USA. Its S-enantiomer, eszopiclone, is approved in the USA.

Zolpidem is rapidly absorbed after oral administration and reaches peak plasma levels in 2–3 h, with 70% bioavailability; ingestion with food slows absorption. Zolpidem has an elimination half-life of approximately 2 h and is oxidized to inactive metabolites. As a result, in most cases, a bedtime dose is completely eliminated during the night, and there is no daytime hangover, predicting usefulness of the immediate release formulation for difficulty falling asleep and early interrupted sleep than for early morning awakening. Zaleplon, which is rapidly absorbed and reaches peak plasma levels 1–2 h after dosing, with only 30% bioavailability due to extensive presystemic metabolism, has an elimination half-life of approximately 1.5 h. Zaleplon is almost entirely metabolized to inactive metabolites. It is primarily useful for difficulty falling asleep and usually does not cause a hangover. Eszopiclone is rapidly absorbed, with an elimination half-life of 5–7 h. About 50–60% is weakly bound to plasma protein. Eszopiclone is extensively metabolized by CYP3A4 and CYP2E1 but does not appear to stimulate or inhibit any CYP450 enzymes.

Adverse Effects
The most common adverse effects of benzodiazepines include sedation, slowed reaction time, and impaired memory and psychomotor function [3]. Especially in older people, benzodiazepines and z-drugs (including zopiclone) can cause falls, syncope, and fractures due to slowed reaction time, impaired balance and gait, and impaired vision [29-31]. Falls are more frequent with chronic benzodiazepine use, long-acting benzodiazepines, and higher benzodiazepine doses [30]. Health-care providers often underestimate the risks of z-drugs compared to benzodiazepines because reports of lower rates of adverse effects have been based on less experience than with benzodiazepines [32]. In doses up to 10 mg, zaleplon has been found to cause dizziness, headache, somnolence, and psychomotor impairment. These side effects usually abate within 2–3 h as a result of the short duration of action of the drug, minimizing daytime impairment. However, at doses of 60 mg, psychomotor impairment persists for 25 h. in addition to potential side effects of all benzodiazepines and z-drugs that are discussed in detail in this section, adverse effects of specific benzodiazepines that have been variably reported are summarized in Table 2 [33-36].

Table 2.
Clinically important adverse effects of specific benzodiazepines

Cognitive Side Effects
Evaluating cognitive effects of benzodiazepines is complicated by the impact on cognition of conditions such as anxiety and depression for which these medications are prescribed [37]. Nevertheless, benzodiazepines have been found on formal testing to impair psychomotor speed, learning, acquisition of new knowledge, visuospatial perception, and immediate memory, an important component of which may be the result of impaired attention secondary to sedation; tolerance often develops to sedation and to its consequent adverse cognitive effects [37].

In a multi-center prospective trial, short-term addition of a benzodiazepine did not impair cognitive function in patients with major depressive disorder, while it may have improved processing speed, possibly by enhancing adherence to the antidepressant regimen resulting in improvement of depression [38]. A European randomized multicenter study of the antihypertensive nilvadipine in 448 patients (mean age 72 years) with mild-to-moderate Alzheimer’s disease (Mini-Mental State Exam score 12–26; median duration of symptoms 3.7 years) reported that use of a benzodiazepine or z-drug (most frequently alprazolam, lorazepam, oxazepam, bromazepam, and zolpidem) did not hasten cognitive decline over 18 months of follow-up [30]. Possible improvement of cognition occurred in a study demonstrating that lorazepam increased a functional magnetic resonance imaging measure of network resiliency in older subjects with poor cognitive scores at the same time that it improved those scores [39]. Older subjects with the lowest cognitive scores had the greatest improvement of cognition with the benzodiazepine, so that they were equivalent to cognitive scores of younger subjects. Any potential benefit for some instances of cognitive impairment seems likely to reflect improvement of attention that was impaired by anxiety rather than a direct benefit of the benzodiazepine.

Although careful testing may reveal persistent dysfunction in multiple cognitive domains after chronic benzodiazepine use in some studies, their impact on daily functioning is usually not apparent [37] and higher cognitive functions are often unaffected [40]. Whether or not benzodiazepines interfere with actual functioning, the risk of increased cognitive impairment caused by sedation warrants minimizing their use in patients with pre-existing impaired cognition caused by neurological factors [41].

Delirium is associated with globally disrupted cortical functioning, altered network connectivity, and/or suppression of reticular activating system activity [42, 43], any of which can be caused by CNS depressants such as benzodiazepines, resulting in inducing or aggravating delirium [30]. In addition, withdrawal from benzodiazepines is a regular cause of delirium in emergency and inpatient settings. Ongoing benzodiazepine use more than doubled the risk of delirium in an adult inpatient study [30], and a dose-related contribution of exposure to benzodiazepines to the risk of delirium independent of other clinical factors was found on a pediatric intensive care unit (ICU) [44]. A systematic review of relevant studies conducted over 10 years (total N = 11,138) found that the majority of the studies identified benzodiazepine use in the ICU as a risk factor for delirium [43]. The possibility of confounding by indication (i.e., patients receiving benzodiazepines may have had illnesses that were more likely to predispose them to delirium) could not be excluded. Use of benzodiazepines has been found to be an independent risk factor for development [45] and duration and severity of delirium [46] in hospitalized children [45]. A component of this risk may be the result of abrupt withdrawal of benzodiazepines that were being taken prior to or shortly after entering the hospital [47].

At one time, intravenous lorazepam was used to treat delirium in critically or terminally ill patients with cancer [48]. However, two Cochrane reviews of the effectiveness and tolerability of benzodiazepines for non-alcohol-related delirium conducted a decade apart [49, 50] found no clear empirical support for use of benzodiazepines for delirium. Given the potential benefits of antipsychotic drugs [51] and electroconvulsive therapy (ECT) [52], the use of benzodiazepines to treat delirium not caused by CNS depressant (including alcohol) withdrawal should generally be considered only when other treatments are ineffective or poorly tolerated.

Benzodiazepines and Dementia
Proposed mechanisms of a postulated association between benzodiazepine use and an increased risk of dementia include decreased cognitive reserve, downregulation of GABAA receptors resulting in reduced resilience of information processing systems, and decreased ability to utilize alternative neural networks because of reduced brain activation [53]. One obvious limitation of any finding supporting this hypothesis is that medications prescribed when dementia is already present cannot be said to cause dementia, and it is not known at what point or for how long in the development of a dementing illness a medication would have to be taken to contribute to the pathophysiology of the disease. In addition, reports of an increased risk of dementia with chronic benzodiazepine use may reflect reverse causality – in other words, benzodiazepines are more frequently prescribed for prodromal symptoms of dementia, such as insomnia due to an altered sleep-wake cycle, anxiety resulting from difficulty keeping track of one’s environment, or depression [54, 55].

Data on whether benzodiazepines increase the risk of dementia have been conflicting [30, 53]. Using claims forms from the Canadian public drug plan, a large case-control study compared benzodiazepine prescriptions 6 years prior to diagnosis in patients over the age of 60 years with a diagnosis of dementia, or with a presumed dementia based on prescription of a cholinesterase inhibitor or memantine, and controls matched for age, sex, and duration of follow-up [54]. Controlling for anxiety, insomnia, and depression, the odds of developing Alzheimer’s disease was increased about 50% after taking benzodiazepines for at least 6 months, the risk being greater with long-acting benzodiazepines taken for longer periods of time. The results were limited by insurance claims submitted by a large number of clinicians without consensus diagnoses, lack of certainty that cholinesterase inhibitors or memantine were actually prescribed for a dementing illness, and a higher incidence of stroke and hypercholesterolemia in dementia patients than in controls unrelated to benzodiazepine use.

A paradoxical finding emerged from a Korean national health-care database retrospective cohort study of a random sample of patients over 50 years of age followed for an average of 5.6 years that compared 1,125,346 benzodiazepine users, 433,429 nonusers, and 17,677 users of SSRIs or serotonin and norepinephrine reuptake inhibitors (SNRIs) [53]. There was a 23% higher risk of dementia in both benzodiazepine and antidepressant users than nonusers, but the longer the duration of benzodiazepine (or antidepressant) use, the lower the risk of dementia, and there was no difference in dementia risk between users of benzodiazepines and antidepressants [53]. A parsimonious explanation is that these medications were prescribed for symptoms of anxiety or depression that accelerated the onset of dementia, lowered the threshold for its expression, or aggravated previously subclinical features of dementia, a process that was ameliorated by the medication. In contrast, a retrospective analysis of data from an open-label randomized trial in 116 family practices in the Netherlands involving 3,526 nondemented people (mean age 74.3 years) living in the community found that after a median of 6.7 years, there was no significant association between incidence of dementia and use of short- or long-acting benzodiazepines in general or in interaction with APOE genotype [55].

Lack of association between the use of benzodiazepines and the development of dementing illness has been reported in other large studies. A Canadian 10-year multicenter study of 10,263 older people living in the community or in institutions found that benzodiazepine use was associated with an increased risk of cognitive impairment (presumably a medication side effect), but not with an increased risk of Alzheimer’s disease or other neurological dementia [56]. An 8-year British prospective follow-up of 8,216 nondemented patients found that anticholinergic drugs but not benzodiazepines were associated with development of dementia [57]. A 10-year Danish cohort and nested case-control registry study of 235,465 adults with a first hospitalization for a mood disorder found no meaningful association between use for any period of time of any benzodiazepine or z-drug and subsequent risk of dementia [58]. In a British nested case-control database study comparing 40,770 people with dementia with 283,933 controls matched on age, sex, history, and other confounders, there was no association of new prescriptions for benzodiazepines/z-drugs and development of dementia 4–20 years later [59]. A PET amyloid uptake study in 268 nondemented older adults found significantly less amyloid uptake, implying lower risk of Alzheimer’s disease, in patients taking benzodiazepines, especially shorter-acting agents [60]. Similarly, in a careful 2-year prospective follow-up of nondemented older adults, compared with benzodiazepine nonusers, benzodiazepine users had reduced beta amyloid expression and no differences in global cognitive function and verbal memory [61].

The bulk of evidence, therefore, suggests that benzodiazepines do not increase the risk of developing dementia. However, benzodiazepines can aggravate cognitive dysfunction in demented patients, who are also more susceptible to delirium caused by benzodiazepine intoxication or withdrawal. While there is no compelling reason to avoid benzodiazepines in otherwise healthy patients for fear of causing dementia, these medications should not be considered initial treatments for anxiety in demented patients.

Impaired Driving
Driving tests, along with neurocognitive batteries revealing clinically relevant effects on executive function, vigilance, and reaction time, suggest that benzodiazepine anxiolytics and hypnotics can cause impaired driving, to which some degree of tolerance may develop after 3 years of treatment [62]. Benzodiazepines were found in 15%, and z-drugs in 13%, of 410 suspected drunk or drugged drivers over the age of 65 years; the most frequently detected medications were zopiclone and diazepam [63]. Similarly, the likelihood of being hospitalized as a result of a motor vehicle crash was increased (OR 5.3) by having been prescribed a benzodiazepine in an Australian sample of 616 individuals aged 60 years and older, especially in men (OR 6.2) [64]. In 72,685 French drivers involved in traffic accidents resulting in injury, the risk of being responsible for the accident was higher in users of benzodiazepine or z-drug hypnotics (OR 1.39), with an even greater risk (OR 2.46) for drivers prescribed more than one pill/day of zolpidem [65]. Zopiclone was not associated with an increased risk of causing an accident in this report, but in a laboratory study comparing the effects on psychomotor tests relevant to driving skills, reaction time was impaired more frequently by zopiclone than alcohol [66].

A meta-analysis of 21 epidemiologic and 69 experimental studies found that benzodiazepines were associated with a 60–80% increased risk of being in a traffic accident and a 40% increased risk of being responsible for an accident, most notably in people younger than age 65 years [67]. Taking benzodiazepines during the day impaired driving independent of drug half-life. Use of diazepam, flurazepam, flunitrazepam, nitrazepam, and zopiclone as night-time hypnotics also impaired driving, at least during the first 2–4 weeks of treatment. Use of alcohol plus a benzodiazepine was associated with a 7.7-fold increased risk of a traffic accident [67]. Other research supports the observation that the combination of a benzodiazepine and alcohol impairs driving significantly more than a benzodiazepine alone [68] – an important point because at least 20% of impaired drivers use alcohol along with another substance [63].

These findings imply that all patients taking benzodiazepines should be warned about the potential for impaired driving, especially in combination with alcohol. Because patients frequently are not aware of driving impairment, it is a good idea either to have their driving assessed in a simulator or ask a family member or friend to ride with them to assess common errors such as slowed reaction time or poor anticipation of changing conditions. The risks to patients themselves and their passengers generally prohibit prescribing benzodiazepines to professional drivers or pilots.

Benzodiazepines are occasionally associated with disinhibited behavior, particularly agitated, impulsive, and aggressive behavior directed toward the self or others [69]. Although there are no formal studies of this issue, risk factors appear to include a past history of aggression or impulsivity, high trait anxiety, hostility, borderline personality disorder, use of alcohol, and higher benzodiazepine doses [69, 70]. The incidence of disinhibition is around 1% and is more likely to occur in children and the elderly [71]. Caution is warranted prescribing benzodiazepines, especially in higher doses, to patients with a history of disinhibited behavior.

Benzodiazepine Interactions
All benzodiazepines have additive sedative side effects with other sedative substances. Additive sedation and potential psychomotor impairment are particularly common with alcohol, other benzodiazepines, sedating antidepressants, opioids, antipsychotic drugs, brexanolone, and valerian. Many benzodiazepines are substrates for CYP3A4 and CYP2C19. Substances that inhibit CYP3A, such as ketoconazole, itraconazole, nefazodone, some calcium channel blockers, some SSRIs, grapefruit juice, and the antiretrovirals lopinavir and ritonavir, or drugs that inhibit CYP2C19, such as fluvoxamine, omeprazole, or ticlopidine, can substantially elevate levels of benzodiazepines, such as midazolam, alprazolam, diazepam, and triazolam, with increased drug effect [72]. However, even though alprazolam is a CYP3A4 substrate, it does not seem to be affected by grapefruit juice because of its high bioavailability [73]. Medications that induce CYP3A, such as carbamazepine, phenobarbital, phenytoin, or rifampin, can decrease plasma benzodiazepine concentrations, reducing the effect of the medication. On the other hand, levels of benzodiazepines primarily eliminated via glucuronidation, such as lorazepam, oxazepam, and temazepam, are not affected by concomitant substances acting on CYP450 enzymes. Given the high level of safety of most benzodiazepines and the lack of a clear correlation between blood level and clinical effect, there is usually no need to avoid medications that act on CYP2C19 or 3A4 unless significant adverse effects or unexpected loss of efficacy occur. Because smoking increases the clearance of benzodiazepines, sedation may increase unexpectedly when patients stop smoking and benzodiazepine levels increase. Examples of important pharmacokinetic and pharmacodynamic interactions of specific benzodiazepines are listed in Table 3 [72, 74-77].

Table 3.
Important interactions of specific benzodiazepines

Benzodiazepine Receptor Antagonists
Flumazenil, which has a benzodiazepine structure, is used to block the actions of benzodiazepines and z-drugs. The most common application of this effect is for the treatment of overdose or to reverse conscious sedation or anesthesia with agents acting on the benzodiazepine receptor. It has also been used experimentally to treat hepatic coma. Flumazenil is administered intravenously because of low bioavailability of oral doses. Because the primary action of alcohol and barbiturates on the GABA receptor complex is not at the benzodiazepine receptor, CNS depression caused by these substances is not antagonized by flumazenil. The same is true of opioids, which act at an entirely different site. Emergency treatment of a known or presumed mixed narcotic/benzodiazepine overdose, therefore, may require administration of an opioid antagonist as well as flumazenil. Benzodiazepine withdrawal may be precipitated by flumazenil, especially in patients taking benzodiazepines chronically. Seizures may occur in patients with epilepsy and in patients who have taken an overdose of tricyclic antidepressants (TCAs).

Clinical Applications of Benzodiazepines and Z-Drugs
Anxiety Disorders
The rapid onset of the anxiolytic effect of benzodiazepines makes them particularly useful in the treatment of acute situational anxiety (technically adjustment disorder with anxiety [78]), especially in settings in which the physiologic consequences of anxiety can be dangerous. While controlled studies are challenging, longstanding clinical experience indicates that rapid reduction of anxiety by benzodiazepines can mitigate the arrhythmogenic effects of increased circulating levels of catecholamines on a damaged cardiac conduction system in anxious or frightened patients with acute myocardial infarction [79]. Benzodiazepines are appropriate for rapid control of anxiety in anticipation of surgery and other procedures or in response to acute injury or illness in children as well as adults [80, 81]. In the absence of any potential return on investment for studies of medications that are off patent, no recent therapeutic studies of benzodiazepines in anxiety have been conducted. Clinical experience suggests that a good response to a benzodiazepine is predicted by acute anxiety with a clear precipitant, previous response to a benzodiazepine, awareness that symptoms are psychological, panic attacks, absence of a history of substance misuse, and request for a benzodiazepine. Acute anxiety, especially in response to a time-limited stress in patients with cardiovascular instability, is preferentially treated with a benzodiazepine.

Many authoritative sources agree that benzodiazepines are effective for chronic anxiety disorders, such as generalized anxiety disorder and panic disorder; however, in recent years, prescribing trends and treatment guidelines have shifted from use of benzodiazepines to antidepressants, particularly SSRIs and SNRIs, based on a belief that benzodiazepines are not as well tolerated and have greater risks of misuse and withdrawal syndromes [82-85]. However, the presumed superiority of antidepressants is not supported by the available evidence [85].

Generalized anxiety disorder studies conducted prior to the introduction of antidepressants reported clear effectiveness of benzodiazepines [5, 6]. Later placebo-controlled trials supported the effectiveness of antidepressants, azapirones, and some anticonvulsants [86]. A network meta-analysis (which attempts to substitute for direct comparisons between treatments by comparing multiple pairs of medication-placebo trials) concluded that benzodiazepines were more effective than antidepressants for generalized anxiety disorder, but less well tolerated [87]. In addition to the likelihood that patients and methodologies in benzodiazepine and antidepressant studies in the meta-analysis were not equivalent because they were conducted at different times with different diagnostic criteria, the inference of better tolerability of antidepressants is contradicted by most other research, such as the finding that benzodiazepines were effective with “mild adverse effects” in a systematic review of 3,753 studies (5 of which were of high quality) of benzodiazepine clinical trials, mostly for generalized anxiety disorder, in older adults, who are more sensitive to most medication side effects [88]. A systematic review of all published medication trials concluded that both benzodiazepines and antidepressants should be considered first-line treatments for generalized anxiety disorder [89].

Early studies of panic disorder included comparisons of benzodiazepines both with placebo and with antidepressants, especially imipramine. A Cochrane review of 24 double-blind randomized controlled trials (RCTs) comparing benzodiazepines to placebo or other treatments in panic disorder in adults with or without agoraphobia (N = 4,233) concluded that the overall likelihood of response was 65% higher with benzodiazepines (RR 1.65; number needed to treat = 4); however, most studies were of poor methodological quality and were at high risk of bias in at least one domain, with a likelihood of unblinding in a number of studies [90]. Studies of antidepressants in panic disorder were similarly judged to be of moderate-to-low quality [91]. Early benzodiazepine studies were conducted with alprazolam in doses up to 6–10 mg/day [92], leading to the impression that this was the preferred benzodiazepine [93]. However, a later meta-analysis of published clinical trials suggested no significant difference between alprazolam and other benzodiazepines at equivalent doses [27]. Direct comparisons of alprazolam and imipramine (with a placebo control) in younger and older adults demonstrated comparable efficacy, at least in short-term treatment [92, 94]. An open-label, randomized comparison of clonazepam (mean dose 1.92 mg) and paroxetine (mean dose 38.4 mg) in 120 patients with panic disorder found that clonazepam had a faster onset of action, but both treatments had equivalent efficacy after 8 weeks [95]. Meta-analyses of separate controlled clinical trials of benzodiazepines and antidepressants have also found no difference in efficacy between the medication classes [96, 97]. A review of 18 studies of TCAs, phenelzine, and newer antidepressants that included a meta-analysis of 11 comparisons of TCAs to benzodiazepines in a total of 2,624 patients with panic disorder with or without agoraphobia found no superiority of antidepressants in anxiety disorders and fewer adverse effects with benzodiazepines than with TCAs [82]. However, higher doses of benzodiazepines may be necessary for panic disorder than other anxiety disorders [92]. A systematic review and meta-analysis of 4- to 12-week randomized placebo controlled studies of acute treatment of panic disorder with a benzodiazepine or an SSRI [98] reported that benzodiazepines and SSRIs both caused somnolence; benzodiazepines were associated with more memory problems, decreased libido, constipation, and dry mouth, but were protective against tachycardia, diaphoresis, fatigue, and insomnia; SSRIs caused more diaphoresis, abnormal ejaculation, fatigue, nausea, diarrhea, and insomnia. Furthermore, a meta-analysis of 56 studies with 12,655 mixed anxiety disorder patients assigned to placebo, antidepressants, or benzodiazepines found that controlling for treatment length and year of study, the effect sizes of SSRIs (Hedges’ g = 0.33) and SNRIs (Hedges’ g = 0.36) were significantly lower than the effect size of benzodiazepines (Hedges’ g = 0.50) [99].

In contrast to a study reporting an equivalent decline in efficacy over time for anxiety disorders of benzodiazepines and antidepressants [99], a 15-month follow-up reported no tolerance to the antipanic effect of either alprazolam or imipramine [92]. In the small number of maintenance studies of patients who responded to an initial 8-week RCT, benzodiazepines and antidepressants remained equivalent in efficacy and safety [100]. In a 3-year extension of the acute open-label comparison of clonazepam and paroxetine cited above [95], treatment with both the benzodiazepine and the SSRI was associated with similar continued reductions in panic attacks and severity of anxiety; clonazepam was associated with fewer adverse effects and a small but significantly greater global improvement [101]. A review and meta-analysis of RCTs of treatment of anxiety disorders and of maintenance studies lasting at least 13 weeks after RCTs [100] found no significant differences between antidepressants and benzodiazepines in any outcome after the initial 8 weeks of treatment, except that benzodiazepines had lower discontinuation rates than placebo. However, controlled prospective maintenance trials that might define the optimal maintenance dose and duration of treatment of chronic anxiety disorders have not emerged [102].

The majority of data, therefore, do not support the contention that antidepressants are superior to benzodiazepines in the treatment of anxiety disorders. Instead, both classes of medication have the potential for benefit, side effects, and withdrawal syndromes [103]. Since all benzodiazepines and most antidepressants in the USA are off patent, there is no motivation by industry to conduct direct comparisons of benzodiazepines and antidepressants to define which patients do best with which treatments, and which patients respond better to combinations of benzodiazepines and antidepressants [84]. In actual practice, since antidepressants and benzodiazepines have equivalent effectiveness and different side effects in the treatment of chronic anxiety, the choice is ultimately up to the patient after being informed both about treatment options and the clinician’s preferences. If rapid relief of anxiety is desirable, benzodiazepines have a faster onset of action. If an antidepressant is chosen, early jitteriness and increased anxiety can be minimized by initial treatment with a benzodiazepine and transitioning to an antidepressant when anxiety abates, or starting with a combination of a benzodiazepine with an antidepressant, gradually withdrawing the benzodiazepine when tolerance develops (usually within a few months) to the activating action of the antidepressant. This strategy is discussed below with the use of benzodiazepines in treating depression. Since most anxiety disorders are chronic or relapsing, chronic treatment is often necessary. Attempts to discontinue a benzodiazepine after many years of a positive response without adverse effects, or to transition to an antidepressant, are often unnecessary and unsuccessful.

Benzodiazepines and z-drugs are particularly useful for insomnia associated with acute stress and hospitalization. These medications may promote sleep by a variety of mechanisms, including alteration of arousal systems in the ventrolateral preoptic area and inhibition of the orexin arousal-promoting system [20]. Benzodiazepines increase sleep duration by 30–90 min, enhance perceived quality of sleep, decrease sleep latency, slightly reduce REM sleep, and suppress slow-wave sleep [20]. Tolerance develops to the hypnotic effect after 1–2 months, and subjective sleep quality decreases after 24 weeks, but efficacy can be retained by intermittent use [20]. In addition, insomnia caused by anxiety may continue to respond to benzodiazepines because tolerance does not usually develop to the anxiolytic effect. Rebound insomnia after discontinuing benzodiazepines has been noted in up to 71% of patients [20].

Flurazepam, temazepam, estazolam, and triazolam have been marketed as hypnotics, but there is no pharmacologic reason why other benzodiazepines would not have similar effects on sleep. The choice of medication depends on the nature of the insomnia. Agents that are higher in lipid solubility, such as flurazepam and triazolam, have a faster onset of action that is useful for initial insomnia, whereas less lipid-soluble agents, such as temazepam, may be more useful for middle or terminal insomnia. Medications with short elimination half-lives, such as estazolam and triazolam, avoid daytime hangover but are not effective for middle and terminal insomnia. Triazolam, which did not increase sleep time significantly in clinical trials, has a combination of high lipid solubility, high potency, and short half-life that can cause withdrawal symptoms, such as automatisms and amnesia, as well as rebound insomnia following a single dose [20]. Flurazepam has a short elimination half-life, but its major active metabolite, desalkylflurazepam, has a half-life longer than 1 day, making it useful for patients with daytime anxiety.

The z-drugs have effects similar to benzodiazepines on total sleep time, sleep latency, number of awakenings, and quality of sleep, without changes in sleep architecture at usual doses, but they may not be particularly beneficial for daytime anxiety. Zolpidem accounts for approximately one-third of all hypnotic prescriptions, having eclipsed benzodiazepines because of its perceived more favorable side-effect profile, although psychomotor impairment has been noted with repeated dosing [104]. Its intermediate lipid solubility and duration of action are useful for early and middle but not terminal insomnia, so an extended-release form (Ambien CR) and a slow-release formulation (Intermezzo) were developed for patients with difficulty remaining asleep. As is true of most other hypnotics, it is usually recommended that zolpidem not be administered for more than 2–8 weeks. This recommendation appears to be contradicted by a study that showed no loss of efficacy over 6 months of treatment with extended-release oral zolpidem [105]. Reduction of sleep latency with zaleplon is similar to triazolam, but sleep time is increased more with zaleplon. Zopiclone and eszopiclone have been studied in adults and elderly patients, in whom it also decreases the number of daytime naps. Because the manufacturer chose to conduct 6-month studies, eszopiclone is approved in the USA for long-term treatment of insomnia.

The American Academy of Sleep Medicine [106] and the European Sleep Research Society [107] agree with expert opinions [20, 108] that the initial treatment of insomnia should be with cognitive behavior therapy (CBT) with sleep hygiene. However, CBT may not be a practical intervention for patients with acute, situational insomnia, especially when it is associated with medical hospitalization. Medications recommended by the American Academy include zaleplon, triazolam, and ramelteon for initial insomnia, and eszopiclone, zolpidem, and temazepam for initial and sleep maintenance insomnia [106]. The Academy did not recommend quazepam, estazolam, or flurazepam because of lack of sufficient data [20]. The European Sleep Research Society endorses benzodiazepines, z-drugs, and “some antidepressants” for short-term (≤4 weeks) treatment of insomnia [107].

Given these considerations, benzodiazepine and z-drug hypnotics should probably be considered first for acute insomnia associated with hospitalization, acute illness, or other time-limited stress. The choice of agent would depend on whether a medication with a longer duration of action would be desirable for sleep maintenance or to ameliorate associated daytime anxiety, or whether daytime or psychomotor impairment associated with these features would be problematic. These medications can also be considered when CBT is not effective or acceptable to the patient with chronic insomnia, for which intermittent dosing may be most effective. Since CNS depressants can depress respiration, benzodiazepines should generally be avoided in patients with sleep apnea, although a relatively high (15 mg) single bedtime dose of zopiclone did not affect sleep apnea parameters in a double-blind crossover study in 28 obstructive sleep apnea patients [109].

Depressive Disorders
The frequent comorbidity of depression and anxiety can reflect co-occurrence of mood and anxiety disorders, or anxiety as a symptom of depression and vice versa [110]. Large clinical trials and meta-analyses, which address aggregate data, do not necessarily guide the treatment of individual patients in whom symptoms of anxiety and depression evolve and interact differently [110]. In addition to the limited experimental data that are available, clinical experience suggests that benzodiazepines can be useful in the treatment of depression, especially when it is accompanied by anxiety.

A systematic review and where possible meta-analysis of 38 randomized placebo-controlled trials of benzodiazepines versus TCAs, or in one study fluvoxamine, in depression with or without anxiety found that benzodiazepines and antidepressants were equivalent in short-term efficacy, without any convincing evidence that benzodiazepines aggravated depression [111]. A Cochrane review that evaluated 10 RCTs (N = 731) conducted through May 2019 of an antidepressant plus a benzodiazepine versus antidepressant monotherapy for patients with major depression found that during the first 4 weeks of treatment, the combination was more effective for reduction of depressive symptoms and for response and remission of major depression, and equivalent for reduction of anxiety [112]. Dropouts for adverse effects were about half as frequent in the combination group. It is not clear whether benzodiazepines might improve mood directly, or indirectly, secondary to improvement of sleep and anxiety.

In patients with anxious depression, initial treatment with a benzodiazepine can produce rapid relief both of anxiety and dysphoria and can block the increase in anxiety that can occur when an antidepressant is added. Combining an antidepressant with a benzodiazepine when therapy is initiated inhibits this phenomenon without interfering with the response to the antidepressant [113], reducing the risk that the patient will discontinue the antidepressant before it is effective [114]. Impressions of poorer outcomes with benzodiazepine-antidepressant combinations are likely to reflect their more frequent use in patients with more severe, complex, comorbid, and treatment-refractory depression that does not respond well to an antidepressant alone, requires more adjunctive treatment, and is associated with more adverse reactions to any treatment [113, 115, 116].

Since anxiety is often initially increased by antidepressants, combining a benzodiazepine with an antidepressant or starting treatment with a benzodiazepine and adding an antidepressant once anxiety abates should be considered for anxious depressed patients, as well as for patients who are not anxious to begin with but who become jittery when an antidepressant is started. In most cases, the benzodiazepine can be gradually withdrawn once antidepressant-induced anxiety abates. While there is evidence that short-term treatment with benzodiazepines can ameliorate unipolar depression, there are no studies indicating that monotherapy with one of these medications prevents relapse or is otherwise effective as maintenance treatment. However, anxiety that persists after depression is fully treated may be an indication for ongoing addition of a benzodiazepine. If a sedating antidepressant does not eliminate insomnia, which may persist for a year or more after complete remission of all other depressive symptoms, a benzodiazepine or z-drug hypnotic should be considered.

Bipolar Disorder
For many years, benzodiazepines, especially clonazepam and lorazepam, were used routinely alone or in combination with an antipsychotic drug or a mood stabilizer for rapid reduction of agitation and insomnia in mania, while the effect of the mood stabilizer was slower to emerge [117]. Data from this era [118] demonstrated that the combination of a benzodiazepine and lithium was as effective as the combination of haloperidol and lithium, with fewer adverse effects. In recent years, the use of benzodiazepines has declined, concomitant with increased use of atypical antipsychotic drugs [119], but it is still appropriate to use oral or parenteral benzodiazepines for mania, either in initial combination with an established mood stabilizer or gradually transitioning to a mood stabilizer as agitation abates. In longer-term treatment, nonspecific anxiety occurring in the context of an affective episode often remits with remission of the episode; for persistent anxiety or insomnia and for comorbid anxiety disorders, valproate, the GABAergic properties of which have anxiolytic effects, or a benzodiazepine would be the preferred agent [120]. Alprazolam should be avoided because it has been reported to destabilize bipolar disorder [121], probably because of its antidepressant properties. Acute treatment of agitation in manic and other psychiatrically ill patients is discussed in detail below. The need for continued use of adjunctive benzodiazepines in bipolar disorder is often a marker of more severe and refractory illness or persistent insomnia rather than poor treatment [122].

A Cochrane review of 34 short-term studies involving 2,657 schizophrenia patients, 8 of the studies comparing a benzodiazepine to placebo addition to an antipsychotic drug and 14 comparing benzodiazepine to antipsychotic monotherapy, found small sample sizes and incomplete data reporting [123]. The only evidence that emerged for benzodiazepines alone or combined with antipsychotic drugs was for brief sedation for acute agitation in schizophrenia patients. A follow-up Cochrane report supported the impression that although they are well-tolerated, benzodiazepines do not augment antipsychotic drugs in the treatment of schizophrenia or improve overall outcomes, their only appropriate use being for short-term treatment of agitation [124]. Using a benzodiazepine for agitation is discussed below.

A meta-analysis of studies published through 2017 found no difference in outcome of schizophrenia between the selective benzodiazepine-1 agonist alpidem, eszopiclone, and placebo [125]. However, one individual study reported superiority of alpidem to placebo in overall improvement of schizophrenia, and one study found that short-term treatment with eszopiclone was effective for insomnia in schizophrenia patients [125]. Benzodiazepines or z-drugs could be considered for insomnia if an increase in the dose of the antipsychotic drug is not indicated to treat psychosis more completely. As discussed in detail below, addition of a benzodiazepine for acute agitation is often preferable to increasing the dose of the antipsychotic drug to make use of the sedative side effect because patients are less likely to adhere to higher antipsychotic doses.

Benzodiazepines are first-line treatments for catatonia. In case series, the remission rate of catatonia with benzodiazepines, frequently lorazepam, has ranged from 60% to 80% [126-128]. A common protocol begins with a lorazepam challenge test of 1–2 mg in adults and 0.5–1 mg in children and geriatric patients, administered orally (including via nasogastric tube) or parenterally [129, 130]. Following a response (≥50% improvement), the dose is increased to 6 mg/day in divided dose. The dose is further increased to 6–16 mg/day, and sometimes to as much as 30 mg/day [130-132]. Oral is less effective than sublingual or IM administration [132]. Another benzodiazepine protocol begins with a 2-mg intravenous dose of lorazepam, repeated 3–5 times per day [133]. The dose is increased to 10–12 mg/day if the initial doses are partially effective. Catatonia in children and adolescents, which appears to be most commonly associated with autism spectrum disorders and other developmental disabilities, neurological and infectious diseases, toxic conditions, and congenital disorders, including inborn errors of metabolism [134, 135], also responds to benzodiazepines [134-136].

At 5 times the lorazepam dose, diazepam has also been found to be effective for catatonia [130, 137]. A combined lorazepam/diazepam approach begins with 2 mg of IM lorazepam [132]. If there is no effect within 2 h, a second 2-mg dose is administered, followed by an intravenous infusion of 10 mg diazepam in 500 mL of normal saline at 1.25 mg/h until catatonia remits. A zolpidem challenge test of 10 mg orally or via nasogastric tube has also been utilized [129]. Response is brief and is usually followed by lorazepam, although zolpidem up to 40 mg/day has been used for ongoing treatment [130].

Several studies suggest that a more robust response of catatonia to lorazepam, at least in lower doses, occurs with a shorter duration of catatonia and the presence of waxy flexibility; factors predicting a poorer response include passivity, mutism, auditory hallucinations describing the patient in the third person, and marked retardation and mutism complicating schizophrenia, especially with chronic negative symptoms [131, 138, 139]. There are no controlled studies of maintenance treatment of catatonia with benzodiazepines, but clinical reports suggest that lorazepam doses in the range of 4–10 mg/day are effective [128], and considerable experience indicates that maintenance lorazepam can be effective in reducing relapse and recurrence [131, 132].

Gastrointestinal Disorders
Patients with unexplained gastrointestinal (GI) symptoms have an increased rate of panic disorder, agoraphobia, and major depression, while many GI disorders are aggravated by anxiety [140]. A number of studies published in the 1970s and 1980s reported improvement of signs and symptoms of peptic ulcer with benzodiazepines [140]. Subsequent work suggested that the role of GI GABA receptors in mediating GI responses to stress and inflammation could mediate the potential benefit of GABAergic medications in some GI disorders [141]. Benzodiazepines are better tolerated than antidepressants by some patients with GI disorders in that they do not cause constipation (like TCAs) or diarrhea (like SSRIs), and they can decrease gastric secretion and induce smooth muscle relaxation [140]. However, controlled studies of benzodiazepines in specific GI disorders have not appeared recently. A benzodiazepine might be considered for anxiety in patients with GI disorders that are sensitive to anxiety, such as irritable bowel syndrome or peptic ulcer.

Cardiovascular Disorders
There is an increased prevalence of anxiety disorders and depression in patients with cardiovascular disease [142]. Reduction of autonomic hyperactivity by benzodiazepines, leading to attenuation of myocardial ischemia and hypertension, can be particularly useful in coronary heart disease [142]. As noted earlier, acute reduction of arousal by benzodiazepines in patients with coronary insufficiency can ameliorate the arrhythmogenic effect of high levels of circulating catecholamines in the presence of a damaged cardiac conduction system [79], making benzodiazepines preferred treatments for anxious patients with cardiac instability.

Alcohol Withdrawal
The standard treatment of alcohol withdrawal syndromes is with benzodiazepines, most frequently lorazepam, diazepam, chlordiazepoxide, or clorazepate [143]. Some protocols use standardized dosing protocols, while others adjust each dose based on the response to the previous dose. Alcohol withdrawal syndromes refractory to benzodiazepines are often treated with addition of phenobarbital, propofol, or the alpha-2 agonist dexmedetomidine [143]. Based on the available evidence, alcohol withdrawal syndromes should be treated initially with benzodiazepines, with consideration of a barbiturate tolerance test and withdrawal protocol if the benzodiazepine is not completely effective [144-146].

As noted earlier, aside from their use in CNS depressant withdrawal, benzodiazepines can be problematic in delirium, which is the most common cause of unexplained agitation in medical and surgical inpatients. In the emergency department and inpatient psychiatry services, agitation is often associated with substance intoxication (especially stimulants), mania, and psychotic and behavioral disorders. Benzodiazepines can be very useful in these conditions.

Since reduction of psychosis with antipsychotic drugs is not immediate, use of these medications to treat agitation associated with mania and schizophrenia relies on their sedative side effects. The goal of rapid sedation without the risks of high or repeated doses of antipsychotics can be achieved with a benzodiazepine. When parenteral medication is necessary, IM lorazepam has been used most frequently, but midazolam, which has a more rapid onset and shorter duration of action, is also available in an IM formulation. Anticipating that an antipsychotic drug will be necessary anyway, most clinicians who use a benzodiazepine for sedation in psychotic patients combine it with an antipsychotic drug when treatment is initiated. This approach allows for lower antipsychotic doses, which can promote long-term adherence. In addition, although the quality of studies is limited, benzodiazepines can reduce acute akathisia caused by antipsychotic drugs [147]. An open-label [148] and a controlled study [149] suggested that the combination of lorazepam and an antipsychotic drug produces more rapid sedation and control of aggression and is better tolerated than antipsychotic monotherapy.

Whether or not psychosis is present, the treatment of agitation in unselected acutely ill patients in the general hospital or emergency department frequently involves an antipsychotic drug alone or in combination with a benzodiazepine. A commonly used IM combination consists of 5 mg haloperidol and 2 mg lorazepam. This tradition is based on a randomized uncontrolled trial that found that the combination of 5 mg of haloperidol and 2 mg of lorazepam produced more rapid sedation than either medication alone, although the need for additional medication or hospitalization did not differ between the combination and either monotherapy [150]. Following a systematic review and consensus discussion of pharmacotherapy for agitation, the American College of Emergency Physicians concluded that lorazepam, clonazepam, and flunitrazepam were as effective as 5 mg of haloperidol, and that 5 mg of IM midazolam produced more rapid sedation than lorazepam or haloperidol, while 15 mg of IM midazolam was more effective than haloperidol plus promethazine for rapid sedation and sleep [151]. In acutely agitated psychiatric patients, 2.5–3 mg of midazolam was found to be a useful first-line treatment, while there was no clear benefit of adding haloperidol to a benzodiazepine [151]. Such findings suggest a change in initial treatment of nonpsychotic agitation from the haloperidol-lorazepam combination to monotherapy with midazolam or lorazepam [151]. Emergency treatment of acute agitation in psychotic or nonpsychotic patients is most effectively treated with midazolam or lorazepam, an antipsychotic drug being added for the treatment of psychosis as agitation abates.

Use of Benzodiazepines in Special Circumstances
Electroconvulsive Therapy
Concern about the anticonvulsant action of benzodiazepines leads many clinicians to avoid benzodiazepines during ECT [152]. An increase in seizure threshold by benzodiazepines appears more likely to interfere with the therapeutic effect of unilateral ECT, which requires stimulus intensities much higher than the seizure threshold than does bilateral placement [153]. A French retrospective study of 70 patients receiving biweekly ECT for major depressive disorder found no difference in number of ECTs, seizure threshold, or seizure duration between patients taking an average benzodiazepine dose of 17 mg/day of diazepam equivalents and patients not taking benzodiazepines when stimulus intensity was twice the seizure threshold with bilateral ECT and 6 times the seizure threshold with unilateral ECT [152]. A double-blind comparison of midazolam versus methohexital anesthesia for ECT found no difference in depression score reduction [154]. Such experiences suggest that if benzodiazepines are to be continued during ECT (for example to reduce anxiety in anticipation of the treatment), they may be less problematic with bilateral than with unilateral ECT, with higher than with lower stimulus intensities, and with lower benzodiazepine doses [152, 153, 155].

Pregnancy and Lactation
Since 1.9–3.1% of pregnant women take a benzodiazepine [156, 157], the question of benzodiazepine use during pregnancy is not trivial. However, data on the topic are contradictory [156]. Reports of associations of adverse outcomes and use of benzodiazepines during pregnancy often do not account for confounding variables, such as the effects of increased hypothalamic-pituitary-adrenal activity in stress-related disorders for which benzodiazepines are prescribed, underreported psychiatric disorders, use of substances including nicotine and alcohol during pregnancy, past history of problem pregnancies (which are more common in women taking benzodiazepines), and other lifestyle and environmental factors [158].

In a Norwegian questionnaire cohort study of 82,030 singleton pregnancies among 69,434 women (mean age 30 years), lower gestational age at birth and higher risk of preterm birth were associated with prenatal use of benzodiazepines or z-drugs [159]. Similarly, after adjusting for publication bias, a systematic review and meta-analysis of 14 studies found significant associations of benzodiazepine use during pregnancy and both gestational age and small for gestational age; an association with neonatal intensive care unit (NICU) admission (OR 2.61) was clinically relevant as well as statistically significant [160]. The possibility that NICU admissions were associated with neonatal benzodiazepine withdrawal was not addressed.

In a meta-analysis of 8 studies, benzodiazepine exposure in utero was not associated with an increased risk of congenital malformations [156]. There was no increased risk of congenital malformations in the 5 studies that specifically examined first-trimester exposure, and no increased risk of cardiovascular malformations in 4 studies that examined this specific domain. However, concurrent use of an antidepressant and a benzodiazepine in the first trimester was associated with an increased risk of major malformations (OR 1.40, p = 0.008). Specific malformations were not reported. A meta-analysis of studies conducted over a 30-year period found no association between fetal benzodiazepine exposure and malformations of the oral cleft in pooled data from cohort studies, but there was a significant association in case-control studies [161]. The contribution of dietary factors and use of substances with known oral cleft teratogenicity to such a correlation [162] is difficult to control in these kinds of large-scale reports.

Evidence of behavioral teratogenicity of benzodiazepines has been mixed. In a sample of nearly 72,000 Norwegian children and their siblings, prenatal exposure to benzodiazepines but not z-drugs was associated with a slightly increased incidence of internalizing behavior at ages 1.5 and 3 years [163]. In contrast, a second Norwegian study of 36,401 children found no significant association between exposure to benzodiazepines and/or z-drugs during pregnancy and internalizing or externalizing behaviors measured on the Child Behavior Checklist at age 5 years [164]. Another Norwegian study of 36,086 children of women with depression/anxiety, sleep disorders, or pain-related disorders, 283 of whom were exposed to benzodiazepines or z-drugs in mid (weeks 17–28) or late (week 29 or later) pregnancy found no increased risk at a mean age of 5.1 years of ADHD symptoms or fine motor deficits with exposure at any time point during pregnancy [165]. The results suggested “no substantial detrimental risk on child fine motor and ADHD symptoms after prenatal benzodiazepine/z-hypnotic exposure alone or in combination with opioids or antidepressants” [165]. Either confounding by indication or higher medication doses in women with anxiety/depression who had late pregnancy benzodiazepine/z-drug exposure could have explained a “moderate association” in this study with gross motor and communication deficits.

The bulk of evidence suggests that the risk of major congenital malformation caused by fetal benzodiazepine exposure is low [166, 167]. In the presence of an indication for treatment with a benzodiazepine during pregnancy, combining it with an antidepressant might be avoided. Use of a longer-acting agent, such as chlordiazepoxide or clonazepam, will reduce the risk of neonatal abstinence syndromes but could contribute to excessive neonatal sedation. Prenatal exposure to shorter-acting benzodiazepines may be associated with neonatal withdrawal.

Benzodiazepines are excreted in breast milk. A Japanese study of 11 nursing mothers and their 1-month-old infants taking benzodiazepines, including alprazolam, brotizolam, clonazepam, flunitrazepam, and lorazepam, among others, found that the ratio of breast milk to maternal plasma benzodiazepine level was consistently less than 1, and the relative infant dose was always less than 10% of the maternal dose except for ethyl loflazepate, which was only slightly greater than 10% [168]. No abnormalities were noted in the infants. The authors concluded that medication exposure is low in infants of mothers taking benzodiazepines. If a benzodiazepine is used during breastfeeding, shorter-acting agents such as lorazepam and oxazepam are preferred [169]. Medication exposure can be further reduced by taking the medication after rather than before breastfeeding. The nursing infant should be monitored for sedation, lethargy, and poor sucking [169]. Withdrawal syndromes can occur in the nursing infant if the mother abruptly discontinues the benzodiazepine.

Elderly Patients
Elderly patients are generally more vulnerable to benzodiazepine effects as a result of increased adipose/lean ratio, resulting in a greater volume of distribution, slowed metabolism, decreased clearance, a longer elimination half-life, and increased steady-state concentrations [3]. In addition, since most benzodiazepines are protein-bound, decreased plasma proteins with age can increase levels of active drug [3]. There is also a greater pharmacodynamic effect as a result of changes in receptor function and signaling [3]. These factors can lead to greater drug effect and accumulation if the dose is not adjusted and monitored, especially when benzodiazepines with longer half-lives are administered more frequently than they are eliminated [3].

Based on altered pharmacokinetics and pharmacodynamics, The American Geriatrics Society recommended that benzodiazepines and related drugs be avoided in geriatric patients, especially those with cognitive impairment, because such factors make the elderly more vulnerable to falls, automobile crashes, and impaired cognition, memory, and psychomotor function [30]. Falls and fractures are a particular concern. For example, in a French study, the risk of falls with injury was more than doubled in patients at least 80 years of age, 9% of falls in this age group having fatal outcomes [170]. A prospective study of 1,285 community-dwelling people aged 65 years and older found that over 1 year, falls recurred in 4.7–46.8% [171]. In a British review of data regarding 27,090 patients with dementia (mean age 83 years) over a 16-year period, compared with patients with insomnia who did not take sleeping pills, risks of fractures, especially hip fractures, falls, and ischemic stroke were increased by 33–96% [172]. Morbidity was similar with benzodiazepines and z-drugs, but mortality rates were lower with z-drugs [172]. Leg fractures, presumably due to falls, have been reported in children but not young adults shortly after beginning benzodiazepine treatment for anxiety disorders [173].

In a survey of 7 European countries and Israel, 14.5–44.1% of 4,156 nursing home residents (mean age 83 years) were taking benzodiazepines and z-drugs [174]. The most frequently prescribed medications were zopiclone, lorazepam, and oxazepam. Monotherapy with z-drugs in France was 21 times higher than the European average. Reports of chronic benzodiazepine and z-drug use by elderly individuals are often interpreted as signs of improper practice; however, some older patients continue to benefit from long-term use without impairment or falls, and they experience intolerable rebound symptoms when attempts are made to withdraw the medication [175]. In such circumstances, benzodiazepines with shorter half-lives and fewer active metabolites, such as lorazepam and oxazepam, are less likely to be problematic than agents with longer half-lives and active metabolites, such as diazepam and chlordiazepoxide [3]. Since it is not known how to predict which geriatric patients will tolerate benzodiazepines [175], caution and careful assessment are necessary when considering such medications in this population.

Benzodiazepines and Suicide
Aside from alprazolam (LD50 300–2,000 mg/kg), the lethality in overdose of benzodiazepines by themselves is low (LD50s in the thousands of mg/kg); but the risk of death with overdose is increased when they are combined with opioids or alcohol [176]. The overall lethality in overdose of z-drugs is generally similar to that of benzodiazepines [177]. However, temazepam, zopiclone, zolpidem, nitrazepam, and flunitrazepam may be more lethal in overdose than diazepam [178, 179]. In 2020, 92.7% of fatal benzodiazepine overdoses also involved opioids, especially illegal forms of fentanyl [180]. In spite of these risks, benzodiazepines and opioids are often prescribed together [181]. In fact, from 1993 to 2014, the US rate of combined use of benzodiazepines and opioids increased from 9.8/100,000 to 62.5/100,000, probably explaining a component of the concomitant increase in benzodiazepine-related deaths [16]. An increase in fatal benzodiazepine overdoses is also attributable to combinations with alcohol, and to an increase in use of nonprescribed (i.e., not obtained from a physician) and illicit benzodiazepines [180].

Is suicide more common in benzodiazepine users? In a Swedish study, 154 psychiatric patients who committed suicide were 1.83 times as likely to have been receiving benzodiazepines as were patients matched for age, sex, and psychiatric diagnosis who did not commit suicide [182]. A 2-year survey found that 20% of suicides had recently taken a benzodiazepine, half of them along with an opioid [183], and a nested case-control study suggested that taking benzodiazepines, especially in higher doses, more than doubled the likelihood of a diagnosis of deliberate self-harm in an emergency department [184]. In contrast, a retrospective case-control study of 6,960 patients with anxiety disorders (2,362 of whom filled benzodiazepine prescriptions) and 6,215 with sleep disorders (1,237 of whom filled benzodiazepine prescriptions) found no significant differences in suicide rate between patients who did or did not use benzodiazepines [185]. Any apparent association of benzodiazepine use and suicide may reflect a greater likelihood that patients who are more anxious or agitated and, therefore, at a higher risk of suicide are prescribed these medications (i.e., confounding by indication). Combining benzodiazepines with alcohol would be expected to reduce the threshold for impulsive behavior, including suicide, but there are no studies of this possibility.

The available data suggest that alprazolam, temazepam, zopiclone, zolpidem, nitrazepam, and flunitrazepam may not be good choices for suicidal patients, while all benzodiazepines and z-drugs are riskier in patients who may combine them with alcohol or opioids. Because fatal benzodiazepine overdoses that do not include alcohol are likely to involve concomitant overdose with an opioid, administering opioid antagonists should be considered if a benzodiazepine overdose does not respond promptly to flumazenil. Suicide risk should be assessed frequently in depressed patients who require benzodiazepines because of high levels of anxiety, which increases the risk of suicide [186].

Issues Complicating Benzodiazepine Use
Tolerance (i.e., increased dose necessary to produce the same effect) to benzodiazepine actions occurs most frequently to sedation, and possibly to psychomotor impairment [187]. As noted in the discussion of treatment of anxiety disorders, tolerance to the anxiolytic effect is uncommon [101]. For example, a 2-year naturalistic follow-up of 204 panic disorder patients, 46% of whom were taking clonazepam alone or in combination with an antidepressant, found that clonazepam produced the same continued improvement as antidepressants without any increase in benzodiazepine dose or worsening of clinical status [188]. Similarly, a database review of all Medicaid recipients in New Jersey during a 39-month period found no clinically significant change in dose over 2 years in the 2,440 patients who were taking benzodiazepines [189]. If dosage escalation does occur, it is usually the result of suboptimal prescribing, attempts to eradicate all symptoms completely, interdose withdrawal and rebound, and poor doctor-patient communication rather than tolerance or attempts to achieve euphoria [190]. With adequate treatment, long-term benzodiazepine treatment is usually associated with maintenance of therapeutic benefit, without a need for dosage escalation [190].

Discontinuation Syndromes
Three syndromes may follow cessation of benzodiazepine receptor agonists. Relapse, the return of the original anxiety disorder or insomnia, begins over weeks to months and follows the chronic or recurrent course of the disorder the medication has been used to treat. Rebound, which is characterized by symptoms of the original disorder that are more intense than before treatment was initiated, typically begins within days to weeks after treatment discontinuation and lasts up to 3 weeks. Rebound anxiety can occur after each dose of a short-acting benzodiazepine, such as alprazolam, and is manifested by interdose anxiety or uneasiness and clock-watching that is often mistaken for drug dependence [191]. Similarly, rebound insomnia on the night following use of a hypnotic is more frequent and severe with short-acting, high-potency hypnotics such as triazolam. Rebound insomnia after discontinuation of long-term benzodiazepine use is compounded by dysregulation of endogenous melatonin rhythms.

Syndromes that involve new features of autonomic arousal that were not present prior to treatment are variably called withdrawal or discontinuation syndromes. Withdrawal begins within hours to days of discontinuation of short-acting benzodiazepines but may not begin for 1–2 weeks after abrupt discontinuation of longer-acting preparations; interdose withdrawal may occur with high-potency, highly lipid-soluble, short half-life benzodiazepines, such as alprazolam [191]. Withdrawal manifested as retrograde amnesia is very common after a single dose of ultra-short-acting benzodiazepines like midazolam. Withdrawal syndromes generally develop after cessation of 4–8 months of treatment with typical benzodiazepine doses or after half that duration of treatment with higher doses. Minor signs and symptoms of benzodiazepine withdrawal include sweating, tachycardia, nausea, myoclonus, restlessness, visual changes, tremor, retrograde amnesia, and confusion. Major withdrawal features include seizures, psychosis, and delirium. Seizures are more common in patients with predisposing factors, such as abnormal EEG or brain damage, and with concomitant use of medications that lower the seizure threshold [191].

Withdrawal from short-acting hypnotics, such as triazolam, can be followed by automatisms at night, with amnesia the next day for the behavior. Withdrawal from bedtime doses of zolpidem, zaleplon, zopiclone, and eszopiclone has been associated in a few reported cases with complex sleep behaviors, including sleep driving [192], resulting in injury or, in a small number of individuals, death from automobile crashes, hypothermia, overdoses, falls, and gunshot wounds [193]. Rebound insomnia is more common than other discontinuation syndromes with eszopiclone.

Acute withdrawal symptoms typically last 2–4 weeks and are more common and severe with high-potency, short to medium half-life benzodiazepines [191]. Attenuated, prolonged withdrawal symptoms have been noted months after discontinuation of long half-life benzodiazepines [194]. Persistent withdrawal syndromes after discontinuation of long-term benzodiazepine use include anxiety, depression, psychosis, cognitive impairment, insomnia, sensory phenomena (e.g., tinnitus, paresthesias, pain, numbness, tingling, vibration, strange skin sensations), motor phenomena (muscle pain, cramps, weakness, tremor, shaking, myoclonus, spasms), and GI symptoms (e.g., food intolerance, abdominal distension) [191]. Such features are more common after long-term treatment and may last 6–12 months, and sometimes years [191, 194]. Such syndromes are rare, but there are no reliable data about their precise prevalence or about factors that may indicate susceptibility to prolonged withdrawal.

Withdrawal after long-term benzodiazepine treatment can be minimized but not always prevented by slow tapering over 3–6 months, and sometimes over years [191]. Once it has developed, withdrawal from benzodiazepines is treated by reinstituting the previous dose and withdrawing the medication more slowly, or by substituting a longer-acting benzodiazepine or another medication acting on the benzodiazepine-GABA receptor complex. Since barbiturates are cross-reactive with benzodiazepines, a barbiturate tolerance test can be used to more precisely determine the dose necessary to suppress withdrawal, after which the barbiturate is gradually withdrawn [144]. Possible adjunctive treatments for benzodiazepine withdrawal include melatonin agonists [195], carbamazepine [196], pregabalin [197], and CBT [198]. There is no literature on management of chronic benzodiazepine withdrawal syndromes [191].

Management of alprazolam discontinuation can be particularly challenging. As chronic doses of alprazolam are reduced below approximately 1–1.5 mg, the benzodiazepine receptor undergoes paradoxical downregulation instead of continuing to upregulate, resulting in intensification of withdrawal symptoms, even with a very slow taper [199]. Some clinicians find that substitution of equivalent doses of clonazepam facilitates withdrawal of the final doses of alprazolam, although this approach is not always reliable because the atypical structure of alprazolam may limit cross-reactivity with other benzodiazepines.

The common belief that benzodiazepines have a high risk of misuse [200] has been challenged repeatedly [3]. An analysis of prescriptions for benzodiazepines over 12 years from the Luxembourg National Health Insurance Registry, which represents approximately 95% of the population and all practicing physicians, found that most people who used benzodiazepines continuously did not escalate the dose [201]. The likelihood of taking high benzodiazepine doses was greater with alprazolam, prazepam, and all benzodiazepine hypnotics, especially triazolam, which could have reflected dosage increases in response to withdrawal and tolerance to the hypnotic effect. A 2015–2016 survey found that 17% of people taking benzodiazepines reported misuse, which was defined as using without a prescription, or not using as prescribed; rates of benzodiazepine misuse were highest in those aged 18–25 years (51%, versus 4% of those aged 65 years or older) [16]. It is not known how many people reporting misuse were taking benzodiazepines prescribed for them versus medications obtained on the street or from friends and family. Amplifying this point, the prevalence of self-reported 12-month nonmedical use (NMU) of benzodiazepines and/or z-drugs in a sample of 10,000 participants representing approximately 53 million adults in the UK was 1.2% [202]. Fewer than 5% of this group reported NMU of both a benzodiazepine and a z-drug, indicating that different populations used each class of medication. The vast majority (70–75%) of nonmedically used benzodiazepines and z-drugs were obtained without a prescription.

The assertion that prescription of benzodiazepines leads to abuse of other substances is not supported by experience [4]. A 2016 national drug adverse event surveillance of 358,247 emergency department visits for harms from NMU of illicit drugs, alcohol, benzodiazepines, and prescription opioids found that although benzodiazepines were involved in 47% of cases, they were the only substance in just 6.5% of cases [203]. In the common setting of misuse of multiple substances, the benzodiazepine is used to augment the “high” produced by other substances rather than serving as a gateway drug [4]. A 2017 survey found that 17% of more than 2 million people being treated in publicly funded substance abuse programs identified benzodiazepines as a secondary or tertiary drug of abuse, whereas only 1% said that benzodiazepines were the primary drugs of abuse [16].

Misuse of nonprescribed benzodiazepines may be becoming more of a problem in younger people. Pediatric benzodiazepine use steadily increased from 2000 to 2015 [204], and in 2018, 3.9% of 10th and 12th grade high school students reported NMU of benzodiazepines [16]. Some adolescents may be switching recreational drug use from illicit drugs to prescription medications obtained on the street or stolen from family members on the grounds that the latter are safer because they are approved by the government. The combination of increased misuse of benzodiazepines and increased distress associated with restrictions due to the COVID-19 pandemic has contributed to an increase in the number of adolescent emergency department visits for benzodiazepine overdoses by 34% with opioids, and by 21% without opioids [204].

A review of published data through 2001 found no empirical support for the recommendation that benzodiazepines should be avoided in patients with a past history of substance misuse [205]. The data indicate that a substantial amount of misuse of benzodiazepines, especially in the context of concomitant use of illicit substances, involves medications not obtained by prescription [202, 206, 207]. In fact, benzodiazepine abuse when the medication is prescribed for appropriate indications is uncommon in patients without a history of a substance use disorder [190]. Despite such data, in September 2020, the US FDA announced that it would revise the boxed warning for benzodiazepines to include warnings about risks of abuse, addiction, dependence, and withdrawal syndromes as well as advice to consider alternative treatments [16]. Similar to the response to triplicate prescriptions, this advice may result in inappropriate withholding of benzodiazepines, and use of cannabis, alcohol, and other substances instead [16]. Caution is warranted with patients with a history of substance misuse, but even in this context, short-term closely monitored use may be appropriate for time-limited acute anxiety in situations described earlier, such as in the context of cardiovascular instability.

reference link :


Please enter your comment!
Please enter your name here

Questo sito usa Akismet per ridurre lo spam. Scopri come i tuoi dati vengono elaborati.